Autonomous vehicles, urban deployment, and welfare effects

This paper provides a theoretical analysis of the effects of autonomous vehicles (AVs) on the spatial structures of future cities. We consider two types of AVs, private AVs (PAVs) and shared AVs (SAVs). We assume that AVs have a lower marginal travel cost than human-driven traditional vehicles (TVs) due to additional utility caused by free activities in AVs, but PAVs have a lower marginal travel cost than SAVs due to better privacy, convenience and comfort. The land released by SAVs due to exemption of parking land occupancy is dedicated to firm production and household residence. We further assume that the government owns the land, regulates the housing development, and rents out the houses to the households. Two urban spatial models are presented and compared: one focusing on TVs and the other on mixed PAVs and SAVs. Both models account for land competition among firm production, household residence, and parking. The effects of traffic congestion on the TV / AV cities are identified through comparing the solutions of the models with and without considering traffic congestion. The finding shows that after introducing AVs, the city size may expand or shrink, depending on the marginal travel cost of AVs and the SAV market share in the AV market. Social welfare may increase or decrease, depending on the fixed cost of AVs, besides AV marginal travel cost and SAV market share. In addition, the total congestion cost may increase, and ignoring traffic congestion effects will cause overestimates of household utility, city size and social welfare.